Statistical Analysis of Metallurgical Mechanical Properties with an Application to Ti-6Al-4V Alloy Fatigue Data.
AIR FORCE INST OF TECH WRIGHT-PATTERSON AFB OH SCHOOL OF SYSTEMS AND LOGISTICS
Pagination or Media Count:
The potential application of statistical methods to metallurgical practices was shown using an example of titanium alloy fatigue data. Materials Laboratory, AFWAL was interested to see 1 whether more meaningful S-N curves could be drawn using statistical methods, 2 which of the data can be identified as outliners, and 3 how the different treatments of the alloy compare. Best-fit curves were determined for each treatments by linear and nonlinear regression analysis. Residual analysis was used to test the assumptions on the random error and to select extreme values for further analysis. Seven of the nine extreme values found did not lie within ninety-nine percent prediction intervals determined about the fitted line and were classified as outliers to be fractographically examined by Materials Laboratory. Regression results of sets of treatments, hypothesized by Materials Laboratory to be drawn from similar populations, were compared to determine whether the treatments within a set agree with the Materials Laboratorys hypothesis. Author
- Metallurgy and Metallography
- Statistics and Probability